Research On Prediction Model Of Abnormal Gas Load Based On Artificial Neural Network

In the actual operation of natural gas,the gas gate station connecting upstream and downstream has become an important transmission and distribution center,which controls the industrial development of upstream and downstream.Therefore,the normal operation of natural gas valve station is particularly critical.This article elaborated on the abnormal flow meter data that occurred in the gate station of a gas company in Inner Mongolia.Using neural network tools,an analysis model of the key factors affecting the abnormality was established,and the flow rate was determined as the key factor.Then,BP neural network and RBF neural The network and the RBF neural network optimized by the ant colony algorithm are used as research tools to predict the load flow of the gate station under abnormal conditions.First,the research process of BP neural network is introduced.Use this research tool to analyze and judge the abnormal conditions of the gate station,determine the three indicators of pipeline temperature,pressure,and flow as input,and judge the abnormal state as output,establish an analysis model of key factors,and the accuracy of model analysis meets the requirements.And based on the analysis,it is judged that the flow rate is the key factor affecting the abnormality.Then,the application principle of RBF neural network and ant colony algorithm are explained in detail,and the process of ant colony algorithm optimizing RBF neural network is explained.In the optimization process,the initial pheromone of the ant colony algorithm is adjusted to give the contribution of a single ant to the overall route.After that,three analysis tools of BP neural network,RBF neural network and RBF neural network optimized by ant colony algorithm were used to establish a prediction model of traffic load under abnormal conditions,obtain the prediction results,and compare the prediction results to prove the optimization The accuracy of RBF neural network model prediction.Finally,the regression model is selected as the research tool to model and analyze the load flow under abnormal conditions,obtain the prediction results,and compare and analyze the prediction results with the neural network model.